cryptolog_58-nsa

DQClO : 40?L9671 / /

THERE'S A NEW WORLD COMING — ARE

YOU READY? (U)

LIP (U)

SYSTEM ACQUISTITION DOCUMENT REVIEW (U)

ANALYSTS OF NSA, ARISE! (U) I

COMING HOME (U) I

CIRC: AN INTELLIGENCE DATA BASE CU! . . 1 ~

BETWEEN THE LINES OF YOUR PERFORMANCE

APPRAISAL fU! . . 1 ~

NSA-CROSTIC No. 30 (U) D.H.W.

NSA/ CSS MILITARY LINGUISTIC PROGRAM 20

Declassified and Approved f<

ar Release by NSA or

10-12-2012 p

ursuant to E.O. 13526,

VI DR Case # 5477$

DOCID: 4019671

top MECieiyr

Published Monthly by PI, Techniques and Standards,
for the Personnel of Operations

VOL. VII, No. 1 - 3

JANUARY — MARCH 1980

PUBLISHER

WILLIAM LUTWINIAK

BOARD OF EDITORS

Editor-in-Chief David H. Williams (1103s)

Special Research Vera R. Filby (7119s)

Traffic Analysis Don Taurone (3573s)

P.L. 86-36

For individual subscriptions
send

name and organizational designator
to: CRYPTOLOG, PI

.. nam-

- -- ,

-- ; • • -•

TOP SECRET

DOCID:

4019671

CONFIDENTIAL

There’s a New World Coming
- Are You Ready? <w

P.L. 86-36

. nalyst response to a changing environ-
, /[ ment is an area that has piqued my

interest for a number of years. This
(u ' interest has become especially pro-
nounced in the last four or five years,
largely as the result of the increase in
the variety and number of automated pro-
cesses that have become available, my
involvement in a number of long-term
research projects, and the prospects of
newer and more sophisticated machine-
based analytic techniques. As a conse-
quence, I have given much thought to the
effects that this increased automation
will have on the analyst and to what we,
as an organization, can do to respond to
the challenges and opportunities that will
arise.

cu) If we are to accurately assess how the
analyst of the future will react to the in-
creased automation of the analytic process,
it is necessary to know the analyst of
today. This is logical, since the majority
of tomorrow's analysts are already on board
and active in the analytic field. It also
becomes essential to define the functions
of the analyst since all personnel who carry
the title "analyst" are not analysts by the
strictest definition of the word. For con-
venience and easy reference I have divided
these "analysts" into three categories which
I've labeled, for lack of any better terms,
"loggers," "case analysts" and "research
analysts."

Jan - Mar 80 * CRYPTOLOG * Page 1

EO 1 . 4 . (c)
P.L. 86-36

CONFIDENTIAL

HANDLE VIA COMINT CIIAMMEL3 ONLY

CONFIDENTIAL

OO&ID: 4019671

(u) The changes expected as we become more
and more automated will alter the methods and
procedures currently used in the analytic
areas. This will have a profound effect on
our operations, but to an even greater extent
it will modify the lifestyle and environment
of each of the three types of analysts I've
defined. Each will be affected in a different
way depending on the analyst's perception of
analysis, his personality and his ability
to adapt.

tu) Once the machine takes over the simple
data manipulation and recording functions,
the "logger" will become, to a great extent,
obsolete. His routine will be upset and his
workaday world completely disrupted. He
will no longer have his logging to fill
his day. He won't know what to do and he'll
be completely lost. For those currently hav-
ing trouble isolating the "loggers" in their
outfits, they'll then be very easy to spot.
Just look for those people sitting with a
blank stare on their faces and for those
making nervous movements at their desks as
they anxiously try to find something to do.
Eventually their names will appear on the list
of those who are abusing their sick leave.

The real "logger" will never recover from
automation. He's lost his place in life.

His day-to-day world will have been destroyed
and he simply won't know what to do. For
many it will be too late to start over. Others
will lack the initiative, while for still
others, it's nothing more than a lack of
talent. The world will have passed them by.
They, of course, will not be to blame. It'll
be the machines' fault, or management's fault,
or maybe just the breaks of the game.

(u) In reality, the "logger" has always been
a clerk with a professional job title and a
professional paycheck. He was created by the
system, and, with the advent of new automated
techniques, will be destroyed by the system.
Some will be salvaged. Some "loggers" will
regroup, retrain and regain a place, perhaps
at the "case analyst" level. Those that can't
make the transition will have to be purged
since the cost of automated systems will call
for a decrease in personnel expenditures.

(in The "case analyst" will survive and
thrive in the new automation. He really
never did care for the logging and data
maintenance functions associated with his
job; therefore, he'll adjust. He'll re-
order his priorities and use the additional
time now available for the development and
analysis that he never could quite get
around to under the old semi-automated method
of operation. His productivity will increase
and mor e technical data on his targets will
result. T

| The ultimate gain will

be the production of more intelligence in-
formation in satisfaction of our requirements.
The "case analyst" will grow as a professional,
honing his skills as he practices his trade,
and in many instances will develop into a
"research analyst."

(u) The "research analyst" will see few
changes in his method of operation as a
result of this new automation. He'll have
to deal with new data bases, new retrieval
programs and new equipment, but his daily
routine will remain pretty much the same as
it is now. He will generally review the
same types and amounts of material and work
on the same types of projects. He should
see some improvements in accuracy and com-
pleteness as a result of the elimination of
the "logger" and the automation and resultant
upgrade of the data bases. He will be able
to spend less time on the verification and data
gathering phases of his assigned tasks and
thereby be able to complete more assignments
in a shorter time. This will allow a more
efficient use of the limited number of "case
analysts" and permit us to achieve maximum
benefit from their talents. Automation will
also improve the morale of the "research anal-
yst" since he will be able to function almost
entirely in his primary capacity.

(u) The benefit of this increased automation
is readily apparent. The talent of our people
can be used to the fullest, with dull, repeti-
tive tasks reduced or eliminated. The timeli-
ness, accuracy and quality of our product will
increase. Personnel not working at their pre-
scribed levels can be eliminated, with a result-
ant savings of money. All of this is, of
course, predicated on manageable machine sys-
tems that will function as designed. Since our
track record for the development of such sys-
tems is not impressive, let's hope that we have
learned from experience, and not attempt to reap
benefits before we have proven follow-on systems.

Jan - Mar 80 * CRYPTOLOG * Page 2

CONFIDENTIAL

HANDLE VIA COMINT CHANNL1L3 ONLY

DOCID: 4019671

UNCLASSIFIED

INTERNATIONAL AFFAIRS INSTITUTE
3rd ANNUAL ESSAY CONTEST

Prizes: 1 st Place $ 1 00.00

2nd Place $ 50.00
3rd Place $ 25.00

Winning entries will also be considered for
presentation in an appropriate Agency
publication

Topic: Related to International Affairs
Enter By: 1 October 1980

To enter (or for more information)
Contact Either :

J B51, Room 3W156
or

07, Room 5 A 1 50 8 36

Winners Announced: October 1980

Jan - Mar 80 * CRYPTOLOG * Page 3

UNCLASSIFIED

DOCID: 4019671

SECRET

EO 1.4. (c)

(U)

PIE,

P.L. 86-36

(U) 3ut why should a linguist have to
spend time identifying languages at all?
Why couldn't an automatic, computerized
method of identifying languages be devel-
oped? To implement such an automatic sys-
tem, we would need to construct a data
base, that is, an information bank where
all of the information needed to identify

- a language could be stored. This data base
could contain models of each langage we
may need to identify. The models would
contain all of the information needed to
identify an unknown text by some compari-
son technique. Thus any automatic lan-
* guage identification scheme would have to
be structured as follows:

■ language models

■ comparison method

<°) For various reasons, it is clear that
automatic dictionaries are the least effi-
cient and least reliable method of identify-
ing languages . 2 Further, linguistic methods
based upon the morphological or syntactic
structure of the language require a huge
amount of effert to construct suitable
models. The only reasonable approach to
language modeling which appears easy to
implement as well as reliable is by statis-
tical methods. It now becomes a matter of
deciding which statistic to use,

< u > We decided to try the simplest approach
first: which method of modeling, which is

inexpensive in terms of storage and CPU time,
will yield acceptable results? Monographic
modeling is indeed cheap, but not very power-
ful. The same can be said for standard Agen-
cy techniques of long standing, such as Delta
I.C. and logweights (see Chart 1).

(U) The next possibility for modeling would
be digraphs. Here is where the solutiog ^as 35.35
found. A digraphic model of a language's '
very easy to construct, requiring a relative 1 ' ^ c '
ly small amount of text as input and very
little storage for the model. A digraphic
model stores the information in a digraphic \
matrix (see Chart 2) in terms of the prob-
ability of each possi ble digraph occuring \

for a given language, f

2 For details see the PI ft .paper of 25 July
1979 bj

P.L. 86-36

Jan - Mar 80 * CRYPTOLOG * Page 4

P.L. 86-36

SECRET

HANDLE VIA COMINT OIIANNELO ONLY

DOCID: 4019671

UNCLASSIFIED

fc.L. 86-36

Jan - Mar 80 * CRYPTOLOG * Page 7

UNCLASSIFIED

DOCID: 4019671

Jan - Mar 80 * CRYPTOLOG * Page 8

P.L. 86-36

DOCID: 4019671

UNCLASSIFIED

WHY ARE ALL THE STREETS IN AMERICA WHICH ARE NAMED AFTER A
FRENCHMAN IN SUCH TERRIBLE CONDITION, AND WHO WAS THIS FELLOW
DETOUR, ANYWAY?

It is generally conceded that book-learning is not the best way
to learn a foreign language. Instead, most people agree that the best
way is to actually live in an environment where the language is present
24 hours a day. When a person sees and hears the words in a foreign
language around him he soon learns what makes sense and what doesn't.

How long, for instance, does it take for even the least language-
oriented person to realize that there isn't any famous person named
Einbahn, after whom so many streets are named throughout Germany and
Austria, all of them coincidentally one-way streets? Well, I'll never
forget the trip that two adults and four children, in one Rambler
station wagon, took from Oberammergau , Germany, to Lake Garda, Italy.

The reason why I'll never forget it is not just the horror I felt as
I was being pushed down the Dolomites (it must have been at 150 miles
an hour) by good-natured Italian truck drivers trying to meet a schedule.
No, the reason is that, during calmer, flatter stretches of the trip,
when I could relax my grip on the wheel and enjoy the scenery, it had
seemed to me that Italian villages have such pretty names. But can
they all have the same name! Hadn't we seen that name a few hours back
when we were crossing that bridge? Then a half-hour ago when we went
past the quarry? And here's anothev village with the same name —

Lavori in Corso. Oh, dammit, I told myself, learning Italian quickly,
it means "Men at Work"!

Right after World War II, as the Cold War was starting up, the
United States had to think of ways to get information to "the Russian
People." The proposed methods included balloons that were supposed to
blow from west to east but often didn't cooperate. One method that was
used at an early stage was to set up an official U.S. radio station,
transmitting in Morse code to all the hams in the Soviet Union. Since
it was an official U.S. radio station, it had to have a callsign be-
ginning with a K (if the transmitter is located west of the Mississippi)
or a W (if located east of the Mississippi) . Since the transmitter was
located on the East Coast, is was a W. And since callsigns often mean
something — WGMS, for Washington's Good Music Station— it was decided
to call this transmitter WRU, for U.S. radio transmitting to RUssia.

But it didn't take long for Washington to learn that the Russian People
didn't believe any of the broadcasts. Eventually someone pointed out
that if you send the callsign in Latin letters, WRU, in Morse code
(.-- .-. ..-), any Russian ham would convert those dits and dahs to the
Russian letters BPy, which, unfortunately, represent the Russian word'
for "I'm lying." Well, back to the balloons...

Excerpted from "Twelve Language Anecdotes in
Search of an Author," by Arthur J. Salemme,
formerly PI, now retired.

To Subscribe to CRYPTOLOG,

Or to Change Your Address, Call
1103s

Jan - Mar 80 * CRYPTOLOG * Page 9

UNCLASSIFIED

SIGNATURE SAO

DOCID: 4019671

ron ornom u s e o w iv

SECURITY CLASSIFICATION

SYSTEM ACQUISITION DOCUMENT REVIEW

COVERNAME

NSA FORM H8274 JUN 79

SYSTEM ACQUISITION DOCUMENT REVIEW

PORTIONS TO WHICH THIS CRITIQUE PERTAINS

(Section) I (Page) I (Paragraph)

(Line)

SAO USE ONLY

CONTROL NO.

ACTION (If marked, indicate retpontlble
individual )

due date (Action completed)

resolved (If marked)

unresolved (If marked, no resolution found
during review, mayfmay not regutre further
attention)

P.L. 86-36

reviewer indicating agreement to final disposition

SAO official final disposition ‘

FORM H8274JUN 79

security classification

Jan - Mar 80 * CRYPTO LOG * Page 10

t o n OF F I CIAL use 0H1 ¥

DOCID: 4019671

ro w off i c i al use oniv

p.L. 86-36

ANALYSTS OF NSA, ARISE!

YOU HAVE NOTHING TO LOSE BUT YOUR PAPER

o This entire article is

LL& TOR OF F ICIAL UOE ONLY

N SA is in the midst of a revolution. Be-
fore anyone rushes to call the Federal
Protective Service, let me quickly add
that the revolution, here at NSA, is an in-
formation services revolution. Webster's New
International Dictionary defines the word re-
volution as a "total or radical change.” A
total change is exactly what is happening, in
a rapidly evolving way, to the methods of
information storage, retrieval, and dissemi-
nation now available or being planned at NSA.
The radical change occurring at this Agency,
of course, represents only a microcosm of the
vast and rapid evolution of computers, data
entry devices, and data storage devices which
is taking place worldwide. New, swiftly
growing industries offering information to
the public, on virtually every subject
known to man, have appeared in the last
decade. This information is now instantly
available by attaching a device about the
size of a portable typewriter, to your tele-
phone.

The New Information Entrepreneurs. T12,:
Information Services, now subscribes to the
data bases of several of these new information
entrepreneurs, including Lockheed's DIALOG,
Bibliographic Retrieval Service's BRS/SDI
Service, and the New York Times ' DATABANK.
Hundreds of hours of manual research time is
now condensed to seconds and minutes when
searching the vast data storage banks avail-
able via a telephone line to computers located
in such places as Palo Alto, California and
Scotia, New York. NSA information science
analysts will query these data banks to aid
in your work-related research on such diverse
topics as chemistry, psychology, economics,
education, physics, engineering, social sci-
ences, medicine, drugs, industry and hundreds
of other subjects. NSA is now considering the
addition of LEXIS and NEXIS services offered
by fyeade Data Central. LEXIS is a legal
citation service covering federal and state
law. NEXIS is an extremely powerful research
tool which provides a full text search capa-
bility of news sources such as the wire ser-
vices, news magazines, newspapers and journals
for a three year period. The addition of NEXIS
to T12's inventory of commercial data bases
would potentially eliminate a great amount of
open source processing now being done.

Government-Sponsored Data Bases. T12
has direct access to a number of gov-
ernment-sponsored data bases which offer
information of great value to NSA analysts.
These include the files of the Defense
Technical Information Center (DTIC), which
provides access to technical reports on re-
search and development projects of the De-
fense Department. These reports are avail-
able in special categories, on microfiche,
in the NSA Library. Other DTIC data bases
| offer descriptive summaries of DoD R § D
! activities. COINS (Community On-line Intel-
ligence System) offers valuable community-
wide intelligence information to NSA custom-
ers with a need-to-know.

NSA's In-House Developments. USA's mis-
sion must, however, be also served by infor-
mation data banks developed and operated in-
house. The useful and valuable classified
information available through the SOLIS system
is, of course, accessible by T12*s information
science analysts in researching your classi-
: fied queries. In the last few years, T12 has
developed and maintained a large number of
its own specialized information data bases
using the M204 language and the IBM 370/168
computer for storage and retrieval. These
machine systems have greatly reduced the
r enormous collections of file cabinets so
familiar to visitors to the old CREF. New
systems, now underway and being planned, will
soon guarantee that the last manual file cabi-
net, like the pterodactyl, will fade into ex-
tinction.

T12 is now hoping to solve one of the great-
est barriers to the rapid conversion of the
written word to digital form by using the
Kurzweil data entry machine. This device,
designed to "read" ordinary print and convert
it to computer -compatible digits, was devel-
oped by the same company which originated
the Kurzweil reading machine for the blind,
now available for use in the NSA Library. The
successful employment of this device, or one
like it will do away with the slow "poking- in”
of data by a human, so essential to the com-
puter storage and retrieval of information.

Through

mation retrieval

1

T12* s in-house infor- P • L

system, analysts will be able
to index the hundreds of thousands of reports

Jan - Mar 80 * CRYPTOLOG * Page 11

FO B BfMCML USE 0N1 V

DOCID: 4019671

TOP 3ECRET UMBRA

which contain information needed by Agency
analysts in their daily work. Storage of
full texts using new ultra-high reduction
techniques on microfiche is being studied
under Pro j ect I I

What This Revolution Will Accomplish,
What does all this mean to the Agency analyst
working against a deadline and seeking infor-
mation vital to the completion of a report?

It means that time- sensitive information will

Manipulation of Electrically Received
Information. Other systems, designed to
handle information received electrically, are
also under developemnt using hardware and

already in use in another Agency

the name given to this pro-

software
system

ject no \l AwAitih^ software modification, will
store large quantities of classified infor-
mation presently received here via teletype
and will enable the researcher to perform
full-time searches by words, subjects, series
numbers and other approaches. Thousands of
documents now being processed manually for
hard copy files, a process which is both
labor intensive and time consuming, will be-
come immediately available to researchers
throughout the Agency using easy-fo-leam
search strategies. Automatic distribution
of these documents, in hard copy, is already
taking place under P roject s

tor hard copy.

will eliminate

r 1

.minate th

and

e need

be available when it still has meaning and
importance. It means that , in many cases,
information will be available to the anal-
yst, 24 hours a day, by querying T12 *s data
bases directly, if required. It means get-
ting an answer to your question in seconds
or minutes instead of hours or days.

The American Revolution freed us, as a
nation, from foreign domination; the infor-
mation services revolution will free, us, at
last, from the overwhelming flood of paper
that has dominated us for so long, stealing
our time and efficiency. It may also, in-
cidentally help save our forests for other
uses.

The next time you call for information
services from T12, don’t be surprised when
your question gets researched and answered
with revolutionary zeal!

* * * *

’ P . L . 86-36

WINS SYDNEY JAFFE AWARD (U)

fTSCl At the 198 0 Crypto - Linguist i c Association Annual Banquet
| of the Language Department of the National
Cryptologic School was presented the Sydney Jaffe Award, the highest
recognition a member of the Agency can achieve in the language field.
His significant accomplishments as an Arabic linguist span an
Agency career of some 24 years. T

■fC) In 1 977 . 1 1 performed pioneer work in developing

National Cryptologic School courses in the Syrian, Iraqi, Libyan and
Egyptian dialects of Arabic. Unique among these is the Libyan
course, the only such course available to the intelligence community.
For this seminal work in course development, he received the Agency ’ s
Meritorious Civilian Service Award.

fUl If there has ever been at NSA a linguistic factotum, it is

| | He yields to no one as a model of the traditions

and ideals set by Dr. Sydney Jaffe.

P.L. 86-36

EO 1 . 4 . (c)
P.L. 36-36

86-36

Jan - Mar 80 * CRYPTOLOG * Page 12

TOP SECRET UMBRA

DOCID: 4019671

UNCLASSIFIED

G

p.

L.

86

This article was originally published in the
WIN [Women in NSA] Newsletter for May 1980.
Since the problem it addresses is certainly
not limited to Agency women, but rather
applies to all employees returning from
tours of duty elsewhere, it is reprinted
here so that it may reach a wider reader-
ship. dhw.

L ast year, I was fortunate to have been selected
for an external training program. I was thrilled,
not only because of the opportunity It provided for me
to learn more about the Inner workings of the entire
Department of Defense, but because 1 felt also that I
would then be assimilated back Into the Agency In a
more responsible position. After all , If the Agency
thought enough of me to have selected me for this
"plum," then would they not want me to use ny know-
ledge, later, to the best advantage?— or so I thought.

Judging from my past job experience In several
areas of the Agency, In addition to my recent schooling,
I felt I was ready to assume a position of greater re-
sponsibility upon my return; however, I was stymied In
ny attempts to find such a position. My two-month at-
tempt to find a job before 1 returned to the Washing-
ton ^rea was greeted with such responses as: "We've

just changed the PNM's, and your name will no longer
be circulated to key components" (as I had been led
to expect prior to my departure); and "We've Just
changed the PW's and you must return to your original
key component, since they sponsored you" (that's been
changed again, 1 understand). (The difficulty was
compounded by my trying to do business via a long-
distance cornnerclal phone.) Rather than being given
a choice of assignments, I was told that I would take
a job In a certain organization* and I drove back to
the Washington area specifically to Interview for this
Job. After the interview, I decided, for several rea-
sons, that this wes not the job for me; however, 1 was
"on the books" as returning to that job, and that was
that, as far as the Agency was concerned. When I re-
turned to the Agency, I personally scouted around and
arranged some Job Interviews, but I felt that I didn't
have the time to pursue potential opportunities be-
cause I was under pressure to be assigned SOMEWHERE.

I finally returned, In a similar capacity, to the group
from which I had left. The other three Agency employees
who were In school with me also returned to the same or
similar jobs. A sad commentary, I feel.

1 am grateful for the opportunity the Agency pro-
vided for me to broaden my knowledge. However, I do
have a few suggestions for both the Agency and for those
employees contemplating applying for external training.
First of all, I know the Agency's reasslmllatlon program
has historically had its problems, and I am confident

that the current administration is doing everything It
can to help remedy the situation. 1 can't speak about
overseas returnees, but It seems to me that those re-
turning from specialized external training, such as the
joint and senior service colleges that provide their
sutdents with a broad background In military affairs,
could be used by the Agency In a variety of areas. I
know that there Is one slot, a one-year tour as execu-
tive assistant to the Deputy Director— that goes to a
returnee from the National War College— but other areas
of the Agertcy &Wld ilso use the special kind of ex-
pertise gleaned from this training. A one-year tour
in any one of these ereas would benefit both the Agency
and the employee. In lieu of that, the employee should
not be "locked into" a job upon his or her return. A
number of Interviews (perhaps three) should be arranged
for the employee. There's no substitute for personal
contact to aid one in deciding if one could effectively
work with an Individual or In a particular area (one
can't make this decision via long distance), and a
specified period of time (a week or ten days) should be
allotted the employee to make up his or her mind.

Falling this, what can YOU do to assure yourself
a good Job upon your return? (This advice holds true
also for those who may be contemplating a change of
jobs.) First, make certain your Personnel Summary Is
up to date, and don't be shy, even before you go on
TDY, about giving It to those who may be In a postil on
to place you. Secondly, decide which areas REALLY In-
terest you; find out what you can about these areas, who
the managers are, and get appointments with them. Host
managers are willing to talk to potential employees.
Thirdly, don't discount the theory of "networking"—

It can work for you. A few of the job Interviews I
lined up were as a result of "contacts" I had made
while working on a group-level staff. A few managers
even remembered me from my stint as a Director's
briefer (and that was In 1973!). In addition, senior
WIN members were very generous In giving me advice and
support.

In short, the opportunities are there; YOU CAN MAKE
IT HAPPEN! .

Now, nine months after my return, I have secured
a very good, career-enhancing position. It’s nice to
have the better job now, but It would have been even
nicer to have obtained It upon my return.

Jan - Mar 80 * CRYPTOLOG * Page 13

UNCLASSIFIED

DOCID: 4019671

CONFIDENTIAL

AN INTELLIGENCE
DATA BASEou

P.L. 86-36

IRC, which stands for Central Infor-
mation Reference and Control, is a
system of scientific and Technical
(S § T) intelligence support which is oper-
ated under the auspices of the Defense In-
telligence Agency (DIA) by the Foreign Tech-
nology Division (FTD) of the Air Force Sys-
tems Command (AFSC). The primary purpose of
CIRC is to support the intelligence infor-
mation needs of the five DoD S S T intelli-
gence productions agencies: the Naval Intel-

ligence Support Center, the Army's Foreign
Science and TEchnology Center (FSTC) , Medical
Intelligence Information Agency (MIIA) and
Missile Intelligence Agency (MIA), and FTD.
Access to the CIRC data base is available to
other organizations, and NSA does have a CIRC
terminal.

CIRC means on-line access to almost 1.5
million documents, and another 3.6 million
are available through batch searching. These
5 million records are available to NSA for
retrospective research through the remote
terminal, which is located in T1232 (Soviet
Data Support Workcenter) , Room 3W032, x5989s.
For on-line searching the terminal converses
directly with the computer located at FTD,
at Wright-Patterson AFB, Dayton, Ohio, and
retrieves document references and abstracts
in response to search queries. CIRC pro-
vides document retrieval in a bibliographic
sense, giving citations to documents re-
lating to a given subject or author, rather
than selected and sorted data ready for in-
sertion into documents which the requester
may be preparing. The intent is to furnish
source material to be studied by the re-
quester, who then has to decide the suit-
ability and validity of the information.

The user does not receive the document it-
self from the computer; instead, represent-
ations of the reference are retrieved, giving

such. things as title, date .of report, report
number or source, text extract, and ending
with a list of specific surnames, facility
names, and nomenclature designators occur-
ring in the document. Should the requester
need the entire document, T1232 can usually
obtain it.

^ A major difference which tends to sep-
arate CIRC from many other data bases is
that CIRC does not use a controlled diction-
ary of terms or keywords. Virtually all the
words in the extracts entered into the CIRC
system are usable in searching, with the
exception of some common words such as "both,"
"who" or latter," which contribute little to
retrieval. Because of the free text type of
search, use of broad terms will certainly
lead to a high chance of irrelevant retrievals
while obtaining all the essential coverage.
Therefore, the requester should be prepared
to screen a great deal of information to avoid
missing anything that seems pertinent. The
other extreme — very narrow terms — will usu-
ally render an output of mostly relevant doc-
uments but eliminating some valid ones in
which the searched items are expressed differ-
ently in the referenced document from the
search terms used — for example, searching
for the SA-7 missile, but not for its nick-
name GRAIL, or looking for the nickname
FISHBED, but not for MIG-21. If SA-7 or
FISHBED were not in the reference, but GRAIL
or MIG-21 were, the reference would not be
retrieved. The indexer can only use terms
that are precisely in the reference and is
not to assume a relationship unless it is
spelled out. Thus, there are problems in
using narrow or specific terms without using
synonyms or related terms to achieve full
subject coverage.

DATA BASE CONTENT. Documents entering the
CIRC data base contain information about jbr-

Jan - Mar 80 * CRYPTOLOG * Page 14

eO NFIDENTIAL

DOCID:

4019671

CONFIDENTIAL

eign science and technology as it relates
to U.S. Array, Navy and Air Force interests.
All scientific disciplines, engineering spec-
ialties and technologies are represented as
available. Worldwide coverage is maintained
with emphasis on Communist countries. U.S.
information is not included except when it is
incidental to foreign data.

00 Domestic and foreign open source pub-
lications are included in CIRC as well as
classified messages, Intelligence Information
Reports and other material of all classifi-
cation levels up to TOP SECRET CODEWORD and
TK. Army, Navy and Air Force finished tech-
nical intelligence publications containing
analyzed information are entered into CIRC,
as well as reports of some non-DoD agencies.
The DoD S § T agencies also participate in
obtaining information from other sources
(such as NSA and CIA which do not submit
reports directly to CIRC) for inclusion in
the CIRC data base.

""W— About 83 percent of CIRC’s on-line data
bases and 93 percent of the total CIRC data
bases — on-line and batch — comes from foreign
literature, 4.7 million out of a total of
5 million references. In 1979 over 1350 sub-
scriptions to worldwide publications were
scanned for content selection for CIRC. This
provides a great number of open source un-
classified references to a broad spectrum
of subjects, nomenclatures, facilities, and
authors or personalities, which are available
for search, again, worldwide in coverage, but
with emphasis on Communist countries. The
other 7 percent of CIRC, nearly 300,000 docu-
ment records, includes finished intelligence
reports. IIR's. sensor reports, and other

(u) Many of the open source references are
not available for computer retrieval else-
where, so this part of CIRC is largely unique.
While the retrieved abstract is in English,
most often the open source referemce will
be in the original language. The reference
in CIRC refers the requester to the original
periodical, journal, newspaper or monograph
for the complete article.

OFF-LINE PRINT ORDERS. Because the on-line
terminal printer is slow, bibliographies are
usually ordered from FTD for off-line print-
ing and mailing to T1232 for forwarding to the
requester. There is only one NSA on-line
terminal; when it is receiving output from
the computer, no searches can be initiated.
Bibliographies classified up to SECRET take
seven to ten calendar days from order to

receipt in T1232. TSC/TK orders take two to
three weeks, since they roust be sent by
courier, and there is only one courier flight
per week from Wright-Patterson AFB to the
Washington area. (If an order misses a
flight, it must wait an entire week until
the next one.) Thus, any request resulting
in a number of retrieve references must allow
ample time for the delievery of the data.

FTD is hopeful that high speed printers will
someday be installed at the principal CIRC
user sites (including NSA) , which will result
in much faster delivery, usually by the next
working day, and will free the remote on-line
terminal for queries. Instead of printing
and shipping by mail or courier, FTD will
use the high speed printers for transmission
of the bibliographies to the system-associated
workcenter. No firm date for this enhancement
is available, but FTD is hoping it will take
place during late FY1980 or early FY1981.

Where possible, T1232 will attempt more ex-
peditious responses for high priority requests
as appropriate. p, L

PROFILES. One additional CIRC feature— the E0
profile — is availble to requesters who want
selected information on a current, recurring
basis, A CIRC profile is a preselected canned
query prepared for a requester arid stored
on-line at FTD. When the CIRC data base is
periodically updated, prior to their being put
, on-line, the newly added documents are auto-
matically compared with the profiles, and
those documents which match the profile search
statements are printed and sent on to the re-
quester. A profile is a good way to maintain
current awareness on a particular topic as
material is added to CIRC. Queries may con-
sist of terms, nomenclatures, personalities,
facilities or organizations, countries, source
(such as a specific publication), or classi-
fication of the references. Queries may be
as narrow or as broad as the requester de-
sires. Profiles may be of any classification.

. 8 6-f36
1.4. (c)

USING CIRC. CIRC usage at NSA has been in-
creasing substantially over the last several
years to the point where NSA is among the
top five users of the data base. This is
despite the fact that the availability of
the data base within the Agency has not been
widely publicized, because, in part, of
computer downtime problems, which have led to
feelings of uncertainty about response times.

Use of CIRC can be cumbersome bacause the
data base is so large it is broken into two
segments: the afternoon, or classified, ses-

sion where a full abstract of the reference
is available for all classified references
(TSC/TK material is available during this
session only) , and the morning, or unclassi-
fied session where a full abstract is avail-
able on-line only for unclassified references.

Jan - Mar 80 * CRYPT0L0G * Page 15

CONFIDENTIAL

DOCID: 4019671

UNCLASSIFIED

If a CONFIDENTIAL or SECRET reference is re-
trieved on-line is the morning session, only
the microfiche number is given. This pro-
cedure is followed because not all terminals
for the unclassified session are in secure
locations, because a number of users have
dial-up terminals which use unsecure tele-
phone lines. Off-line bibliographies do
print the complete abastracts of classified
morning references.

The morning unclassified session is so
large that FTD's IBM 360/65 cannot contain
all of the references, so it has been broken
into three segments:

■ CIRC, which consists of the CONFI-

DENTIAL and SECRET material, as
well as the last two years of un-
classified input. This is avail-
able on-line every morning except
Wednesday.

■ CIIO, which contains unclassified

material entered into the system
within the past two to five years.
This is available on Wednesday
mornings.

■ ARCH, consisting of: 'the archival data

base of unclassified references which
were input over five years previously.
This material is available only for
batch searches.

Thus a requester wanting complete coverage
from all of CIRC might receive four separate

LIP (Continued from page 8)

outputs, all containing different references,
from CIRC, CIIO, ARCH, and from SISA, the
TSC/TK data base. Of course, searches can
be limited by classification or by date span.
The requester can limit his retrievals by
indicating specific desires when making his
request.

(u) The primary costs to NSA for the use of
CIRC are the expenses of the 24-hour dedi-
cated circuit to FTD, paper for the printer,
the cost of the Teletype Model -40 Keyboard
Display Printer, and the salary of the one
analyst who spends most of her time doing
CIRC-related activities. There is no charge
for search time, printing or mailing, ot
for CIRC microfiche.

(0) All in all, CIRC is an excellent data
retrieval system, which is to a considerable
extent unduplicated by any other system.

* * * * * * * *

Analysts who have not been aware of
the existence of CIRC, and who might
have questions concerning the system,
or who might wish further information
on how CIRC retrievals of profiles
could help them, should call the NSA
CIRC remote terminal location, T1232,
on 5989s, or drop in at Room 3W032,
where T1232 personnel will attempt
to answer their questions.

00 An interactive computer program has been
developed on the LODESTAR system for use by
the analyst/linguist. The program uses the
technique discussed above, and has complete
instructions for its use contained in the
program itself. The program can usually
identify the correct language in under three
seconds of CPU time. For further deails,
contact one of the authors.

Editor's note: Since the writing of this

article, the data base has been expanded by
the inclusion of the major languages of Europe.

cu) The authors feel that this "distancing"
tecnnique described here probably has other
applications in addition to language identi-
fication. Readers who wish to discuss such
applications are requested to call either Dr.

SOLUTION TO NSA-CROSTIC No. 29

"[A Propos ed Cure for the] Tim e-in-Grade
Syndrome," I [CRYPTOLOG,

November 197/ .

"In general, promotion decisions [must]
ultimately determine [just] who will make
the key decisions within [an] organization.
...The initiative for promotion must belong
to the managers and it is their responsi-
bility to see that worthy personnel are
promoted. It [must] be the employees' re-
sponsibility to qualify for promotion."

P.L.

86-36

Jan - Mar 80 * CRYPTOLOG * Page 16

UNCLASSIFIED